Prompting 101: Writing Better Instructions for AI Tools

By TalentQuill Team

The quality of what you get from AI tools is largely determined by the quality of what you put in. Here's a practical framework for writing prompts that work.

Prompting is the most immediately learnable AI skill. Unlike critical evaluation, which takes time to develop through practice, prompting can be improved significantly in a short period with a few structural principles.

Be specific about the output you want

Vague prompts produce vague outputs. Instead of "write a summary of this document," try "write a 3-paragraph executive summary of this document, focusing on the key recommendations and their financial implications." The more precisely you describe what you need — format, length, focus, tone, audience — the more useful the output.

Provide context

AI models don't know who you are, what you're working on, or what you're going to do with the output unless you tell them. "I'm a product manager at a B2B SaaS company preparing for a quarterly review. Our Q2 retention rate dropped by 4 percentage points. Help me structure a root-cause analysis." That's dramatically more useful than "help me understand why retention dropped."

Ask for reasoning, not just conclusions

If you ask AI to make a recommendation, it will. If you ask it to explain its reasoning, identify the key uncertainties, and note the assumptions it's making, you get something you can actually evaluate and build on. "What are the three strongest arguments on each side of this decision, and which do you think is most important and why?" gets you further than "what should I do?"

Iterate

Good prompting is rarely a single exchange. It's a conversation: you get an output, identify what's missing or wrong, and refine. "That's helpful, but I need you to be more specific about the implementation timeline" or "the tone is too formal — can you rewrite this for a casual Slack message?" This turns AI into a genuine creative collaborator rather than a one-shot tool.

One more thing worth remembering: prompts that work well are worth saving. If you've developed a prompting approach that consistently produces good outputs for a particular task — analysing customer feedback, drafting emails to stakeholders, summarising research papers — keep it. Build a personal library of prompts that you can return to and adapt.